Cell, Vol. 65. 605616,
May 31, 1991, Copyright
0 1991 by Cell Press
Tn7 Transposition In Vitro Proceeds
through an Excised Transposon Intermediate
Generated by Staggered Breaks in DNA
Roland Bainton, Pascal Gamas,* and Nancy L. Craig
Department of Biochemistry and Biophysics
and Department of Microbiology and Immunology
George W. Hooper Foundation
University of California, San Francisco
San Francisco, California 94143
Summary
We have developed
a ceil-free system In which the
bacterial transposon Tn7 inserts at high frequency into
its preferred target site in the Escherichia coli chromosome, aftTrt7; Tn7 transposition
in vitro requires ATP
and Tn7-encoded
proteins. Tn7 transposes
via a cut
and paste mechanism in which the element Is excised
from the donor DNA by staggered double-strand
breaks
and then Inserted Into aft7117 by the joining of 3’
transposon
ends to 5’ target ends. Neither recombination intermediates
nor products are observed in the
absence of any protein component
or DNA substrate.
Thus, we suggest that Tn7 transposition
occurs In a
nucleoproteln
complex containing
several proteins
and the substrate DNAs and that recognition
of attTn7
within this complex provokes strand cleavages at the
Tn7 ends.
Introduction
Mobile DNA segments translocate by a variety of mechanisms. These mechanisms differ in the nature of the recombining sites, the type and order of strand breakage
and joining reactions, and the degree of involvement of
DNA replication. We are interested in understanding
the
molecular basis of transposition, a reaction in which a discrete DNA segment moves between nonhomologous
DNA
sites (Berg and Howe, 1989). This reaction is distinguished
by the fact that it i?volves three distinct DNA segments,
the two ends of the transposon and the target site; no
sequence similarity between the transposon ends and the
target site is required. Another hallmark of transposition is
that element insertion is accompanied by the duplication
of several base pairs of target sequence. This duplication
results from the introduction and subsequent repair of a
staggered break in the target DNA, the transposon being
joined to the ends of the overhanging target strands at this
break.
We are particularly interested in understanding the trans
position of the bacterial transposon Tn7 (Barth et al., 1978;
Craig, 1989), because this element displays an unusual
degree of insertion specificity. Most transposable
elements show relatively little insertion site selectivity when
transposing to large DNA molecules (Berg and Howe,
‘Current Address: Laboratoire
de Biologie Moleculaire
Plantes-Microorganismes,
CNRS-INRA,
BP27-31326
TOLOSAN
CEDEX, France
des Relations
CASTANET
1989). By contrast, Tn7 inserts at high frequency into a
single site in the Escherichia coli chromosome called
attTn7(Barth et al., 1978; Lichtenstein and Brenner, 1982;
Craig, 1989). Tn7 is unrelated in nucleotide sequence to
affTn7 (Lichtenstein and Brenner, 1982; Gay et al., 1986;
McKown et al., 1988); thus, no DNA sequence homology
is involved in the selection of affTn7 as a preferential target. This element’s transposition machinery is elaborate.
Tn7 encodes a surprising array of transposition genes,
rnsABCDE. Insertion into aflTn7 in vivo requires four of
these genes, fnsA6C + tn.sD (Rogers et al., 1986; Waddell
and Craig, 1988; Kubo and Craig, 1990). Large DNA sites
at both ends of Tn7 (Arciszewska et al., 1989) and at attTn7
(McKown et al., 1988; Gringauz et al., 1988; Qadri et al.,
1989) are also required. When attTn7 is unavailable, Tn7
resembles most other transposable elements, moving at
low frequency and inserting into many different sites (Rogers et al., 1986; Waddell and Craig, 1988; Kubo and Craig,
1990). This alternative transposition pathway requires a
distinct, but overlapping,
ensemble of four tns genes,
tnsA6C + tnsE.
Although many transposable elements have been characterized genetically, only a few have been examined biochemically (Mizuuchi, 1983; Brown et al., 1987; Morisato
and Kleckner, 1987; Eichinger and Boeke, 1988). The high
frequency of Tn7 insertion into attTn7 suggested that this
reaction would be amenable to biochemical dissection.
We report here the development of a cell-free system for
Tn7 transposition. We find that ATP and four tns fractions,
i.e., protein preparations derived from cells containing individual tns genes, are required for the formation of recombination intermediates or products; no recombination
is
detected when any fraction is omitted. Furthermore, the
presence of the atfTn7 target site is required for the initiation of recombination, i.e., the production of recombination
intermediates. These results suggest that Tn7 transposition is executed by a nucleoprotein assembly containing
multiple proteins and the substrate DNAs and that the
proper assembly of this entire recombination machine is
required for the strand breakage reactions that initiate recombination.
We also demonstrate that Tn7 transposition in vitro is a
nonreplicative
reaction that proceeds by a series of
double-strand breaks that disconnect the transposon from
the flanking donor DNA to produce an excised transposon,
which is then inserted intoaftTn7. The DNA breakage reactions at the transposon ends and at the target DNA both
occur by staggered breaks that generate 5’ overhanging
ends. Tn7 insertion occurs via single-strand joins between
the 3’ transposon ends and the 5’ ends of target DNA.
Results
Establishment
of a Cell-Free System for Tn7
Transposition
Tn7 transposition in vitro occurs efficiently when substrate
DNAs, a donor plasmid containing a mini-Tn7 transposon,
DSB.L
DSB.R
..
INTERMEDIATES
TARGET
Figure
SIMPLE
INSERTION
1. Tn7 Transposition
Substrates,
Intermediates,
and Products
The 6.1 kb donor mofecule contains a 1.6 kb mini-Tn7 element flanked by sequences
unrelated to affTn7; the mini-Tn7 element contains segments
from the left (Tn7L) and right (Tn7R) ends of Tn7 that provide the cis-acting transposition
sequences.
The 2.9 kb target molecule contains a 0.15
kb affTn7 segment flanked by EcoRl sites. Double-strand
breaks at the junction of either Tn7L or Tn7R with the donor backbone
(arrows) produce
the DSB.L and DSBR species, which are transposition
intermediates.
An excised linear tranaposon
(ELT) is generated by a second double-strand
break in the DSBs. The excised mini-Tn7 element inserts site- and orientation-specifically
into attTn7, producing a 4.5 kb simple insertion product
and a 4.5 kb gapped donor backbone.
EcoRl digestion of the simple insertion product releases a diagnostic
1.75 kb fragment.
Restriction
sites:
E = EcoRI, P - Pstl. and S = Seal.
and a target plasmid containing aftTn7(Figure l), are incubated at 8Cl°C with tns protein fractions, ATP, and MgAc.
Recombination is detected by extraction of DNA from the
reaction mixture, digestion with restriction enzymes, separation by gel electrophoresis,
and hybridization with specific probes.
After incubation, several new species are detectable
with a mini-Tn7-specific probe (Figure 2A): a simple insertion transposition product (mini-Tn7 in affTn7) and several
recombination intermediates, includinqdonor
DNA molecules that are broken by a double-strand break at either the
Tn7L-backbone
junction (DSB.L) or the Tn7R-backbone
junction (DSSR) (see Figure 1). Insertion specificity was
verified by DNA sequence analysis of cloned affTnZ:mini-
Tn7 segments. All insertions were orientation-specific,
i.e., with Tn7R oriented towards the bacterial g/mS gene,
and were accompanied
by a target duplication of 5 bp.
More than 85% of the insertions occurred at either of two
adjacent nucleotide positions in affTn7, and a few nearby
insertions were also observed. This distribution closely
resembles the pattern of Tn7 insertions into affTn7 in vivo
(Liihtenstein and Brenner, 1982; Gay et al., 1988; McKown
et al., 1988; Gringauz et al., 1988). The efficiency of the
cell-free reaction is striking, especially in view of the fact
that it is conducted in crude extracts: translocation of
about 25% of the Tn7 elements from donor molecules to
target molecules is readily observed.
Tn7 transposition to ahTn7 in vivo requires four Tn7-
Figure
BACKBONE
:
L
1
2
3
4
5
6
7
S
9
10
4
-,
INSERTION+
1
2
L
2. Tn7 Transposition
In Vitro
(A) RequirementsforTn7transposition.
Shown
is an autoradllram
that displays
EcoRIdigested
species
detected
with a mini-Tn7specific probe from tranaposf6on
reactions
in
which the indicated components
were omitted.
DNA mofecules
are lab&d
as in Figure 1;
DSB.L and DSB.R appear very different in size
because
of the EcoRl diitffn.
Reactffns
were performed
as described
in the Experimental Procedures,
except for lane 6, where
pfasmld lacking attTn7 was used as a target
instead of the usual aftTn7containing
plasmfd.
Asterisks denote that the indicated Tns protein
was supplied in a crude fraction that also contained host proteins. Following the preincubation step, reactions were incubated for 5 min at
3D°C, except that shown in lane I, which was
stopped immediately
after mixing of the reaction components.
Lane L shows DNA size
markers.
(8) A gapped donor backbone
is a transposition
product. An ethidium-stained
agarose gel is shown, which displays the EcoRIdigested
substrates
(lane 1) and products (lane 2) of an in vitro Tn7 transposition
reaction; DNA molecules
are labeled as in Figure 1. After recombination,
two new
species are evident: an attTn7 segment molecule containing a simple insertion of mini-Tn7 and a donor backbone from which the mini-Tn7 element
has been excised. Lane L shows DNA size markers.
Tn7 Transposition
SO7
In Vitro
encoded genes, tnsABC + tnsD (Rogers et al., 1986; Waddell and Craig, 1968; Kubo and Craig, 1990). Tn7 transposition in vitro can occur efficiently when four protein
fractions, each derived from cells containing one of these
fns genes, are mixed (Figure 2A). In this experiment, the
tnsA fraction is a slightly fractionated extract containing
TnsA, the tnsB fraction is substantially purified TnsB protein, the rnsC fraction is a slightly fractionated extract containing TnsC, and the tnsD fraction is partially purified
TnsD protein. It is notable that no transposition intermediates or products are detected when any one of the tns
fractions is omitted (Figure 2A, lanes 3-6).
We also find that no recombination
intermediates or
products are detected when attTn7 is omitted (Figure 2A,
lane 8). This result is particularly interesting, as it suggests
that the transposon ends “see” the target site before recombination actually begins, i.e., before any double-strand
breaks occur at the transposon ends: We have been unable to detect any changes in the target plasmid when any
tns fraction or the mini-Tn7 donor plasmid is omitted from
the reaction (data not shown).
These experiments show that the four tns fractions and
the attTn7 target are required for recombination.
Notably,
the production of both recombination
intermediates and
products requires the presence of all the reaction components, i.e., no partial reactions are detected when any
single component is omitted. An attractive hypothesis that
explains these results is that Tn7 transposition is a highly
concerted reaction carried out by a nucleoprotein assembly (machine) containing the DNA substrates and multiple
recombination proteins. Because transposition in vitro requires combining four different fractions, at least four different proteins are likely to participate directly in recombination.
Tn7 transposition
in vitro has several other requirements. ATP (Figure 2A, lane 10) or dATP (data not shown)
must be added. Efficient recombination
also requires 10
mM MgAc (lane 9). If MgAc is included at the beginning of
the incubation, little recombination is observed (data not
shown); substantially more (perhaps 50-fold) recombination occurs when MgAc is added after a preincubation
step. Transposition is most efficient with supercoiled substrates, but considerable transposition can be observed
with relaxed DNAs (data not shown). Recombination
is
stimulated by the presence of polyvinyl alcohol or polyethylene glycol (data not shown).
Tn7 Transposition
In Vitro Does Not Involve
Replication
of the Element
A key issue in understanding
a transposition reaction is
determining whether the transposon is copied by DNA replication or whether it is transferred from the donor molecule
to the target molecule via a nonreplicative, cut and paste
reaction. Several observations indicate that Tn7 transposition in vitro is nonreplicative. In addition to the simple insertion of minLTn7 into affTn7, another transposition product
is a gapped donor molecule, i.e., a donor backbone from
which mini-Tn7 has been excised (Figure 28). The simple
insertion product and gapped donor backbone appear with
approximately
the same time course and in equimolar
amounts (data not shown). Production of this gapped donor backbone indicates that Tn7 translocation occurs by
cutting the transposon away from the donor DNA and joining it to the target DNA without replication, a view supported by the demonstration that an excised transposon
is a transposition intermediate (see below).
Another indication of the nonreplicative nature of Tn7
transposition is that recombination in vitro is not affected
by the inclusion of DNA synthesis inhibitors such as dideoxynucleotides
under conditions in which a strong inhibition of DNA polymerase activity is observed (data not
shown).
Double-Strand
Breaks at the Ends of Tn7 Produce
Recombination
Intermediates
Several types of evidence indicate that the DSBs, i.e.,
donor DNA molecules broken by a double-strand break at
a transposon end, are indeed transposition intermediates.
These species have the kinetic properties of intermediates: they are most abundant at early times and disappear
at late times during the incubation (Figure 3). These kinetic
properties are particularly evident when the overall rate of
transposition is slowed by lowering the incubation temperature to 22%. Here, the DSB species are detected earlier
and in greater amounts than are the transposition products; it is also notable that a considerable fraction of the
substrate donor DNA is converted to DSBs. Disappearance of the DSBs is not due to nonspecific nuclease degradation, because a control fragment was not degraded in
these reactions. The view that the DSBs are transposition
intermediates is also supported by the fact that they, like
the simple insertion product, are generated by staggered
breaks in DNA (see below). Similar levels of DSBs are
observed when the reactions are stopped by a variety of
Figure 3. The Double-Strand
Break Species
ties of Transposition
Intermediates
*
Donor
*
+
*
InPenlo”
0SB.R
“SOL
Have the Kinetic
Product
Proper-
Shown is an autoradiogram
that displays &al-digested
species detected by a mini-Tn7-specific
probe produced in transposition
reactions, after incubation for the indicated times at either 30% or 22OC
(preincubation
also at 22%). Reactions were performed as described
in the Experimental
Procedures,
except that they also contained
a
1200 bp linear fragment recognized
by the probe (“Control”) to evaluate
nuclease degradation.
DNAs are as described
in Figure 1; the DSBs
appear to be different in size because of the Seal-digestion.
Cdl
808
Figure 4. An Excised Transposon
Is a Recombination Intermediate
(A) Production of excised transposon.
Shown is
an autorwram
that dtsp#ays Ecof-diied
spectes ds&cted with a mini-Tn7-specMc
prcbe
added NaCl
from transposition
reactiins that contained var(mM)
0
25 50100200
ious amounts of monovafent
satt. Lanes l-5:
reactions supplemented
with the indited
salt;
‘..
the salt concentrations
supplied by protein frac+
DONOR
6.1
tions were 30 mM KCI and 3 mM NaCI. Lane
L: DNA size markers with a prominent 1.6 kb
5.1
species. DNAs are as described in Figure 1.
4.1
(B) An “excised”
tmnsposo n is an effective
transposttion
substrate. Shown is an autoradiogram of the EcoRIdiited
products detected
with a mini-Tn7-speciflc
probe of transposition
reactions containing about 25 ng (0.25 nM) of
Ml&generated
‘excised”
transposon
as the
mini-Tn7 substrate, instead of the usual miniTn7-containing
donor plasmid. Reactions were
carried out as described
in the Experimental
Procedures,
lacking particular components
as
indicated. Asterisks denote that the indicated
L 1
2
3 45
1
2
3 4
5
6 7
8
Tns protein was supplied in a crude fraction
that also contained host proteins. Addition of
crude extract (CE) from cells lacking Tns proteins stimulates, but is not essential for, recombination. DNAs are labeled as in Figure 1, except that the “excised” transposon
is generated
by restrfction
instead of recombination
(as in
Figure 4A and elsewhere).
(C) Mlul transposon.
An Mlul transposon-containing plasmid was produced as described in
the Experimental
Procedures.
Digestion with Mlul enzyme, which cuts at the indicated positions (arrows), reteases a linear *excised’ transposon, whose 3’ ends are cleanly exposed (as also with authentic excised transposon
generated in recombination
reactions) and whose 5’ ends
have 4 nt extensions (as opposed to the 3 nt extensions
of authentic excised transposon).
methods,
(data not
tein-DNA
provoked
including treatment with proteinase and phenol
8hOWn); thus, the DSBs are not likely to be procomplexes, nor is DNA breakage likely to be
by our manipulations.
Exclssd Tn7 Can Bs a Transposition
Substrate
In addition to DSB8, a transooson species that has been
excised from the donor backbone by double-strand breaks
at both transposon end8 is produced in the incubation
mixtures. This excised specie8 is present in all reactions,
but often at a very low level. However, high levels of excised transposon and low levels of insertion product are
evident under 8ome reaction conditions, for example, with
high NaCl (Figure 4A). We have been unable to observe
disappearance
of the excised transposon, a8 might be
expected for a recombination intermediate. However, high
levels of excised transposon may reflect substantial impairment of the insertion step. Like the simple insertion
product and the DSB species, the excised transposon is
generated by staggered breaks in DNA (see below).
To determine whether an excised mini-Tn7 can participate in recombination,
we asked whether exogenously
supplied, “excised” mini-Tn7 could insert into atiTn7when
present a8 the sole transposon Substrate. We introduced
restriction sites at the Tn7 termini such that restriction
enzyme digestion generates a mini-Tn7 species with staggered ends (Figure 4C), the 3’ ends being cut adjacent to
Ll and Rl and the 5’ end8 cut within the flanking DNA to
generate 4 nucleotide (nt) overhangs. (As described below, we believe that the excised mini-Tn7 species generated during transposition has exposed 3’ ends and 3 nt
overhangs on its 5’ ends.) The exogenously supplied miniTn7 element also has two sequence Change8 in its Tn7L
segment.
The restriction-generated
“excised” mini-Tn7 inserts efficiently in site- and orientation-specific
fashion into affTn7
(Figure 48) supporting the view that the excised transposon is indeed a transposition intermediate. As with a plasmid 8UbStrate, insertion of the “excised” transposon requires the presence of all of the tns fractions, ATP, and
MgAc. In addition to demonstrating that a transposon speties that is not connected to a donor backbone can participate in transposition, these results suggest that no 81s
function is involved exclusively in the excision stage of
recombination.
We also note that we have been unable to detect a strand
transfer intermediate under any condition, either in complete Tn7 transposition reactions or when any tns component is omitted. This species, in which the transposon, the
donor backbone, and the target are covalently linked, is a
key intermediate in bacteriophage Mu transposition (Shapiro, 1979; Craigie and Mizuuchi, 1995; Pato, 1989).
The Polarity of Strand Brsakags and Joining
Tn7 insertion is accompanied by the duplication of 5 bp of
target sequence (Lichtenstein and Brenner, 1982), pre-
Tg
Transposition
In Vitro
A
,60
4
T"7.L
Figure
B
5’ OVERLAP
5. Transposition
Can Occur
269
103
.,,T",.R
via Two Different
Polarities
of Strand
,57
attTn,-R
T"7.L
Breakage
and Joining
The duplication
of target sequences
upon transposon
insertion is presumed
to result from cleavage of the target site by a staggered break and
subsequent
repair of the gap resulting from linkageof
the transposon
ends to the target site ends. Two polarities of target cleavage and transposon
joining can be imagined: (A) the target is cleaved by a staggered
break generating 5’ overhanging
ends that join to the 3’ transposon
ends, or (6)
the target is cleaved by a staggered break generating 3’overhanging
ends that join to the 5’transposon
ends. The two polarities can be distinguished
by identifying which target strands are covalently joined to which transposon
strands in the simple insertion transposition
product. Asterisks
mark
the specific, diagnostic gaps in the transposition
product. The primers Tn7-L, attTn7-L, Tn7-R, and attTn7-R are complementary
to particular strands
in affTn7 and the Tn7 ends. Extension from these primers yields products of different and defined lengths as indicated (thick arrows), depending
on which polarity of target breakage and strand joining is used in Tn7 transposition.
Restriction
sites: E = EcoRI, P = Pstl.
sumed to result from the cleavage and subsequent repair
of a staggered break in the target DNA (Grindley and Sherratt, 1979; Shapiro, 1979). Two alternative types of target
breaks can be imagined (Figure 5), one producing B’over-
A
hanging ends (Figure
hanging ends (Figure
age will result in two
transposon ends and
5A) and the other producing 3’over58). These two types of target cleavdistinct types of joints between the
the target site: either the 3’transpo-
Figure
Target
B
6. The
Ends
3’ Ends
of Tn7 Join to the 5’
Shown are autoradiograms
of DNA sequencing gels on which the products of primer extensions are displayed,
in which the simple inserP
+269
tion transposition
product was a template with
'- +232
+213
the indicated primers. Also shown schematically are the relevant Tn7 end and affTn7 regions (see also Figure 5). The lengths of the
extension
products were determined
by comCl30
parison with dideoxy sequencing
ladders. The
lengths (26 or 30 nt) and mobilities of the primers are also shown.
(A) The 3’ transposon
ends are covalently
joined to attTn7. Extensions
from the target
attTn7-L
and attTn7-R
primers
using the
EcoRCdigested
and Pstl-digested
simple insertion product as templates are shown. The observed products (horizontal arrows), 213 nt for
attTn7-L
and 269 nt for attTn7-R,
are the
lengths expected
if the 3’transposon
ends are
joined to attTn7 (Figure 5A). No product is evi(57)
dent at the positions corresponding
to free 3’
transposon
ends, extension from attTn7-L is 34
30
nt, and that from attTn7-R is 57 nt.
;
(B) The 5’ transposon
ends are not joined to
c
attTn7. Extensions
from the transposon
end
Tn7-L and Tn7-R primers using Pstl-digested
simple insertion product as template (P: left lane of pair) and, as a control, an intact affTn7::mini.Tn7
fragment as a template (C: right lane of pair).
Two classes of products are observed from each primer with the simple insertion template. One class of products extends to about the Ytransposon
ends (black arrows), approximately
60 nt from Tn7-L and 130 nt from Tn7-R (the exact lengths of these products are considered
in Figure 7). Thus,
some 5’ transposon
ends in the simple insertion product are not linked to affTn7, corresponding
to a 5’ staggered
break in the target (Figure 5A).
The other class of products (upper, gray arrows) extend to the positions expected for linkage of the 5’transposon
ends to the target (103 nt from
Tn7-L and 232 nt from Tn7-R), probably reflecting DNA repair.
h
Cdl
810
son ends are joined to the 5’ target ends (Figure 5A), or
the 5’ transposon ends are joined to the 3’ target ends
(Figure 56). In the first case (Figure 5A), gaps between the
5’ transposon ends and the 3’ target ends are expected;
the second breakage and joining polarity leaves gaps adjacent to the 3’ transposon ends (Figure 58).
To determine the polarity of Tn7 strand breakage and
joining, we determined the positions of the intact and
gapped strands in the simple insertion product, by using
it as a template for extensions with end-labeled oligonucleotide primers specific to the transposon ends and the flanking afUn DNA.
Our analysis has revealed that the target site is broken
by staggered breaks with 5’overhanging
ends (Figure 5A).
Primers attTn7-L and attTn7-R anneal with the strands of
the target DNA that, following transposition using 5’target
breaks, will be joined with the 3’ ends of the inserted
transposon. We find that the attTn7-L extension product
is 213 nt and that the attTn7-R product is 269 nt (Figure
6A). These lengths correspond to the distance from the
primers to restriction sites within the transposon ends(Figure 5A). Thus, in the simple insertion product, these template strands are intact, i.e., the 3’ transposon ends are
covalently linked to the Vends of the target. Similar primer
extension analysis of the gapped donor backbone supports the view that Tn7 excision involves cleavages directly adjacent to the 3’ transposon ends, which are then
joined to target DNA (data not shown).
If the 3’ ends of the transposon are joined to the target
DNA, i.e., are intact, then it is expected that the other
strands are gapped (Figure 5A). The state of the 5’transposon ends was examined by extension from the Tn7-L and
Tn7-R primers (Figure 68). For both primers, two distinct
classes of extension products are observed. From Tn7-L,
there is a cluster of products at about 60 nt and another
species at 103 nt from Tn7-R, there is a cluster of products
at about 130 nt and another species at 232 nt. The shorter
class of extension products from each primer (those at
about 60 nt for Tn7-L and about 130 nt for Tn7-R) correspond closely to the distance to the 5’ transposon ends.
(The exact lengths of these products are considered in
detail below.) These results demonstrate that a considerable fraction of the 5’ transposon strands are broken and
thus are not covalently linked to target DNA.
The longer extensions from the Tn7-L and Tn7-R primers do extend to lengths consistent with covalent linkage
of the B’transposon end with the target DNA, i.e., for Tnir-L,
103 nt and for Tn7-R, 232 nt. We hypothesize that these
species result from some repair in this crude system of the
gap that initially flanks the 5’ ends. Repair of the right end
gap appears to be more efficient than repair of the left end
gap, i.e., more 232 nt Tn7-R product is observed than 103
nt Tn7-L product.
We interpret these results to indicate that during Tn7
transposition, the target DNA is cleaved by a staggered
break that yields 5’ overhanging ends and that the 5’ ends
of the target DNA are joined to the 3’ ends of the transposon (Figure 5A). We presume that a 5 bp gap flanks the 5
transposon ends and is subsequently converted to duplex
form through DNA repair.
Staggered
Breaks at the Transpoeon
Ends
What type of DNA strand cleavages release Tn7 from the
donor molecule? Analysis of the simple insertion product
suggests that Tn7 is cut away from the donor DNA via
staggered breaks, in which one DNA strand is cleaved
precisely at the 3’transposon end, and the other strand is
cleaved at a displaced position within the flanking donor
DNA. We were led to this view through our characterization
of the 5’ ends of the transposon in the simple insertion
product by extension from the primers Tn7-L and Tn7-R
(see Figure 5A). If the 5’ ends were cut flush with the 3’
ends, i.e., at exactly the end of the transposon, we would
expect the Tn7-L extension product to be 60 nt and the
Tn7-R product to be 130 nt. Rather, we find that the Tn7-L
extension product (Figure 7A) and the Tn7-R extension
product (Figure 78) are both 3 nt longer than expected.
This result would be obtained if cleavage of the 5’ strands
occurs within the flanking donor DNA.
If Tn7 is excised from the donor backbone by staggered
breaks involving strand cleavage within the flanking donor
DNA, we would expect sequences from the flanking donor
backbone to be attached to the B’transposon ends. Chemical sequence analysis of the 63 nt Tn7-L extension product
reveals that the nucleotides attached to the 5’ terminus of
the left end of Tn7 are those adjacent to this transposon
end in the flanking donor DNA (data not shown). This result
provides strong evidence that cleavage of the 5’ transposon strand actually occurs at a position displaced from
the transposon end via a staggered break. We have also
observed such staggered breaks at the transposon ends
when Tn7 transposes in vitro from another unrelated donor
site (data not shown); thus, these staggered breaks are
not peculiar to a particular donor site.
We have performed similar primer extension analysis of
the ends of the DSBs and the excised transposon generated in in vitro transpositkm reactions (Figure 7). In all cases,
the structure of the 5’ transposon strands was identical,
i.e., 3 nt of donor backbone areattached to the transposon.
These results support the view that the DSBs and the excised transposon are authentic transposition intermediates and that the transposon is cut away from the donor
backbone by staggered breaks in DNA (Figure 6).
The donor sequences attached to the excised transposon are not essential for transposition.
An “excised
transposon species generated by the polymerase chain
reaction (PCR) and primers whose 5’ strands end with the
transposon termini (as in Figure 7, lane 1) inserts efficiently
into aftTn7 (data not shown).
Tn7 Targst Immunity Is Active In the Cell-Free
Tmnspositlon
System
Tn7 displays target immunity in vivo; that is, the presence
of a copy of Tn7 in a target DNA molecule substantially
reduces the frequency of a second Tn7 insertion into that
target (Hauer and Shapiro, 1964; Arciszewska et al.,
1969). Target immunity is fundamentally related to transposition, in that both events require the same transposon
end sequences (Lee et al., 1963; Adzuma and Mizuuchi,
1966). The presence in a target molecule of the transposition sequences from the right end of Tn7 (Rl-199) is suffi-
Tn7 Transposition
811
In Vitro
DONOR
BREAKS
+TFlANSPOSITION
PRODUCT I 63 nts
+
TRANSPOSON
END I 60 ntr
TARGET
BREAKS
INITIAL
PRODUCT
Donor
4
5
Tn7L
FINAL
PRODUCT’
Figure 8. Strand Breakage and Joining during Tn7 Transposition
We suggest that both the ends of Tn7 and attTn7 are broken by staggered breaks that generate 5’overhanging
ends. Breaks at the target
site are proposed to be 5 bp in length, and those at the ends of Tn7
are proposed to be 3 bp in length. The 3’ Tn7 ends join to the 5’
target site ends, generating an initial transposition
product whose 5’
transposon ends are attached to several nucleotides of donor DNA and
are flanked by 5 bp gaps. DNA repair fills in the gaps and concomitantly
removes the nonhomologous
donor sequences from the 5’transposon
ends, generating an intact duplex flanked by 5 bp target duplications.
The target sequence shown is the 5 bp in aftTn7 usually duplicated
upon Tn7 insertion.
; Donor
Figure 7. Donor Nucleotides
Remain Attached to the 5’ Transposon
Strands in Transyition
Intermediates
and the Simple Insertion
Product
Shown are autoradiograms
of DNA sequencing gels that display primer
extensions from Tn7end oligonucleotides
(see Figure 5A), under conditions where nontemplated
addition does not occur (see Expsrimental Procedures).
The templates
were: lane 1, as a controt, a PCRgenerated
tranaposon
that terminates
exactly at the 5’ transposon
ends; lane 2. excised linear transposon (ELT); lane 3, the DSB species;
and lane 4, the simpte insertttn transposition
product (Sl). At the left
is a dideoxynucleotkte
sequencing
ladder generated with the same
oligonucteobdeon
an intactaffTnir::mini-Tn7fragment.
The nucleotide
sequence of the corresponding
S’transposon
strand and the sequence
of the flanking donor strand in pEM are shown, along with the distance
in nucleottkfw
to the ottgonucteotkte
primer. At the bottom, the positions of strand &avages
at the tranaposon termini are shown schematically; the 5’ deavage
positions determined
in this experiment
are
indicated with an asterisk.
(A) Extension from the Tn7-L primer. Extension using the simple insertion as template yields a product 3 nt longer than expected for extension to the 5’ transposon
end.
(8) Extension from the Tn7-R primer. Extension using the simple insertion as template yields a product 3 nt longer than expected for extension to the 5’ transposon
end.
cient to provide a high level of immunity in vivo; by contrast,
a truncated right end (M-42)
cannot provide immunity
and is also defective in transposition (Arciszewska et al.,
1989).
We find that Tn7 target immunity is operative in the cellfree transposition system. We evaluated immunity in vitro
by examining the target activity of attTn7 plasmids that
also contain Tn7R segments (Figure 9). In the experiments
shown in Figure 9, lanes 2 and 3, the reactions contained
a mixture of two target DNAs, one an attTn7plasmid (target
A) and the other a plasmid containing both an atiTn7segment and a segment from Tn7R, either Rl-199 or R1-42
(target8
B and C, respectively), located at a distance of
about 1 kb from attTn7. No mini-Tn7 insertion into the
attTn7segment of target B is detectable (Figure 9, lane 2),
although insertion into the control affTn7 molecule target
A is observed. This result suggests that the presence of
RI -199 in target B inactivates this molecule. In contrast,
the presence of Rl-42 in target C does not block insertion
(Figure 9, lane 3). We note that in the presence of target
B, a modest decrease in insertion into the control target A
Target
Target
A = attTn7
B = sttTn7 & RI-199
Target
C = attTn7
+
-
& W-42
insertion
DONOR
into B
insertion
into
6.1
-
kb
++
+ -
+
+
C
e
e
7.1
6.1
f
5.1
f
4.1
f
3.1
e ,2.0
insertion
into
A
1.75
kb -b
e .1.6
123L
Figure 9. Tn7 Transposition
in Vitro Displays Target Immunity
Shown is an autoradiogram
that displays EcoRI-diQested
species detected with a mini-Tn7-specific
probe from transposition
reactions in
whkhtheaff7n7tar~etmoleculealsocontainedaTn7Rse~mentabout
1 kbfrom affTn7. Target A is theaffTrr7target
used in other experiments
(pKAD4-3);
Target B@lAl 1) and Target C (pLA24) contain affTn7and
the indicated Tn7R segment. The concentration
of the target plasmids
(3 nM each) is about Mold less than that of the standard conditions,
to avoid a trans-inhibition
by Tn7RcontaininQ
targets (see text). The
positii
of a simple insertion product in each target plasmid is also
shown. Lane 1, target A; lane 2: target A + target B; lane 3: target A
+ target c.
molecule is also obsenred (compare lanes 1 and 2). We
suspect that this decrease results from a competition between the Rl-199 segment in target B and the donor DNA
for transposition proteins that interact with the transposon
ends (the molar ratio of donor to target 6 is 1:7); such an
effect has been observed in vivo (Arciszewska et al., 1989).
Discussion
We have developed a ceil-free system that faithfully reproduces many aspects of Tn7 transposition in vivo. Tn7 inserts in vitro at high frequency in site- and orientationspecific fashion into its preferred target site, attTn7.
Transposition
occurs efficiently when exogenous DNA
substrates, one a donor molecule containing a mini-Tn7
element and the other an atiTn7containing
target molecule, are incubated with extracts derived from cells containing the Tn7-encoded transposition
genes tnsABC +
fnsD required for insertion into aftTn7 in vivo (Rogers et
al., 1988; Waddeli and Craig, 1988; Kubo and Craig, 1990).
The in vitro system also displays the phenomenon of transposition immunity that Tn7 displays in vivo; the presence
of Tn7 in a target molecule blocks subsequent insertion of
another copy of Tn7 into that target molecule (Hauer and
Shapiro, 1984; Arciszewska et al., 1989). Tn7 transposi-
tion in vitro also requires the presence of ATP, which is
likely to be directly involved in recombination,
because
TnsC is an ATP-binding protein (P. Gamas and N. L. C.,
unpublished data). An auxiliary role of ATP may be to facilitate supercoiling of the DNA substrates by promoting the
action of DNA gyrase (Gellert et al., 1978).
Tn7 Transposition
C+~rs in a Multiprotein
Complex
An attractive hypothesis is that the four Tns proteins participate directly in recombination.
Transposition in vitro requires the mixing of four protein fractionsderived
from four
different bacterial strains, each containing one of the four
tns genes, fnsABC + tnsD. Specific roles for some Tns
proteins have emerged from other experiments (see below). It is likely that host proteins will also be directly involved; most transactions involving extrachromosomai
DNA elements are mediated by a combination of elementand host-encoded proteins (Pato, 1989; Thompson and
Landy, 1989; Kleckner, 1990). Mutations in several E. coii
genes can affect Tn7 transposition
(0. Hughes and
N. L. C., unpublished data). Reconstitution of Tn7 transposition in vitro with purified proteins will be required to identify both the Tns- and host-encoded participants in recombination.
We propose that the Tn7 recombination
proteins, together with the substrate donor and target DNA% form a
occurs. Elaborate protein-nucleic
acid complexes also
mediate other recombination react-,
DNA replication,
transcription, and RNA processing (Echols, 1990). We are
attracted to the view that Tn7 transposition occurs in such
a complex, because no transposition
intermediates
or
products are detected in the absence of any required protein or DNA component. Thus, we imagine that the complete transposition machinery must properly juxtapose the
three DNA segments involved in recombination-the
two
transposon ends and the target DNA-and
the multiple
transposition proteins, prior to the initiation of the strand
breakage and joining reactions that underlie recombination. In this scenario, atiTn7 is recognized by the rest of
the recombination
machinery before strand breakage at
the transposon ends occurs. Thus, specific recognition of
attTn7 actually provokes transposition by stimulating the
initiation of recombination.
What are the roles of the recombination proteins? Other
work from this laboratory has established that two of the
Tns proteins are sequence-specific
DNA-binding proteins:
TnsB binds to the ends of Tn7 (McKown et al., 1987; L.
Arcisewska, R. McKown, and N. L. C., unpublished data),
and TnsD binds to afUn (Waddell and Craig, 1989; K.
Kubo and N. L. C., unpublished data). Possible roles for
TnsA and TnsC (and/or host proteins) include acting as
linker proteins to mediate the specific juxtaposition of the
TnsB-bound transposon ends and the TnsD-bound target
site and interacting with the positions of strand breakage
at the in7 termini and at attTn7.
Tn7 Transposition
Is Nonreplicative
Two distinct pathways for the movement of transposable
elements have been identified: nonrepiicative transposi-
Tn7 Transposition
013
In Vitro
tion, in which the transposon is cut out of the donor site and
pasted into the target site; and replicative transposition, in
which a copy of the transposon is made by DNA synthesis
and inserted into the target site (Berg and Howe, 1989).
We have established that Tn7 translocates in vitro via a
nonreplicative
mechanism (Figure 1). The principal evidence for this view is that the transposition products we
observe are a simple insertion of Tn7 into altTn7 and a
linear donor molecule from which Tn7 has been excised.
In vivo studies also provide evidence that Tn7 transposes
to attTn7via a nonreplicative mechanism (K. Orle, R. Gallagher, and N. L. C., unpublished data).
The fate in vivo of the donor backbone from which Tn7
is excised is unknown. Tn7 transposition in vivo provokes
expression of the SOS system (A. Stellwagen and N. L. C.,
unpublished
data), consistent with the production of a
gapped donor backbone (Roberts and Kleckner, 1988).
We do not observe simple rejoining of the backbone in
vitro or in vivo (data not shown). Perhaps the broken donor
backbone is lost when transposition occurs, i.e., donor suicide occurs. Another possibility is that the broken donor
molecule is repaired by a double-strand-gap
repair reaction using an unbroken Tn7-containing donor molecule as
a template, as has been observed in Drosophila P-element
transposition (Engels et al., 1990).
The Tn7 Transposition
Pathway
Our analyses have revealed that Tn7 transposition occurs
by a series of double-strand breaks and single-strand joins
(Figures 1 and 8). Double-strand breaks at the transposon
ends produce a linear excised transposon that is subsequently inserted by single-strand joins into an atrTn7target
site that is also broken by a double-strand break.
Double-strand cleavages of a donor molecule at both
Tn7L and Tn7R generate an excised transposon intermediate. These double-strand breaks are staggered, generating 5’overhanging ends such that the S’transposon termini
are cleanly exposed, but several nucleotides of donor
backbone remain attached to the 5’ transposon strands.
We observe that three donor nucleotides are attached to
the 5’ transposon ends, suggesting that these staggered
cleavages are 3 bp in length. Because the species we have
analyzed were generated in a crude system, however, it
is difficult to exclude the possibility that a species with
longer donor sequences is initially produced and then exonucleolytically reduced. Nonetheless, our studies have established that the transposon ends are released from the
donor backbone via staggered breaks. The presence of
donor sequences on the 5’ transposon strands is not essential for transposition.
The double-strand breaks at the termini of Tn7 need not
occur simultaneously.
Similar amounts of donor molecules broken at the junction of the donor backbone are
observed with either Tn7L or Tn7R, which are both converted to insertion product. The ends of Tn7 are structurally and functionally distinct (Arciszewska et al., 1989).
The orientation-specific
insertion of Tn7 into atiTn7 also
implies discrimination between the transposon ends. Our
observations here suggest that discrimination
between
Tn7L and Tn7R occurs at a step other than formation and
utilization of the double-strand
break species.
The target site must also be broken by a staggered
double-strand break; the joining of the ends of this break
to the transposon ends and subsequent repair of the resulting gaps result in the hallmark target sequence duplication that accompanies transposon insertion (Grindley
and Sherratt, 1979; Shapiro, 1979). Like the cleavages
at the transposon ends, the staggered target cleavage
generates 5’ overhangs; the target cleavage overhangs
are 5 bp in length, as opposed to apparently 3 bp at the
transposon ends. In contrast with the donor DNA, no cleavage of the target DNA is detected prior to its linkage to the
transposon ends.
Upon Tn7 insertion, the 3’ transposon ends are ligated
to the 5’target ends (Figure 8). Formation of the new bonds
between the target site and the transposon termini involves single-strand joins in which high energy phosphodiester bonds must be formed. The energy source for the
covalent bonds between transposon ends and target
strands is as yet undefined in Tn7 transposition. However,
we have established that the energy of the transposondonor bonds is not essential to the formation of the transposon-target
bonds: a transposon lacking the transposon-donor bonds can join to the target site.
It will be important to identify the catalytic centers that
execute strand cleavages at the transposon ends and at
the target site. Because the two types of breaks have the
same polarity, i.e., staggered with 5’overlaps, both cleavages may be carried out by the same polypeptide with
position specificity provided by other proteins that bind
specifically to the transposon ends and to aftTn7. Alternatively, the end and target cleavages may be carried out by
different polypeptides with their own specificity determinants. Since conversion of the excised transposon to insertion product apparently requires the same tns fractions
as a reaction using an intact donor molecule, no fns function appears to act exclusively in excision.
Comparison
of Tn7 with Other
Transposable
Elements
Tn7 transposition resembles the movement of retrotransposons, particularly in forming an elaborate nucleoprotein
recombination complex. The origins of these complexes
are quite different: the Tn7 DNA in its complex is formed
by excision, whereas retrotransposon
DNA in its complex
is formed by reverse transcription of an RNA intermediate
(Bowerman et al., 1989; Varmus and Brown, 1989). However, the initial joints between the element ends and insertion sites are very similar. For both elements, the 3’
transposon strands are linked to the 5’ target ends, and
several nucleotides that will not appear in the final transposition product are attached to the ends of the 5’transposon
strands (Fujiwara and Mizuuchi, 1988; Brown et al., 1989;
Eichinger and Boeke, 1990). With Tn7, these extra nucleotides derive from staggered breaks at the transposon termini; with retrotransposons, these nucleotides derive from
single-strand cleavages of the linear retroviral DNA. These
extra nucleotides are presumably removed during repair
of the gaps that flank the initial insertion product.
Cdl
014
Prior to linkage with the insertion site, the phosphodiester bonds that link Tn7 to its donor site are broken. Tn7 is
excised from the donor backbone by double-strand breaks,
and formation of a strand transfer intermediate is not observed. The strand transfer intermediate, i.e., a joint molecule in which the transposon, donor, and target are linked,
is a key intermediate in both nonreplicative and replicative
bacteriophage
Mu transposition (Shapiro, 1979; Craigie
and Mizuuchi, 1985; Pato, 1989). Tn7, however, appears
to move exclusively by a nonreplicative reaction; excision
of the transposon prior to linkage to the target effectively
precludes a replicative reaction. Although distinct in temporal order, the strand cleavages at the ends of Mu and
Tn7 are chemically similar, as are their target cleavages.
In both systems, the 3’ transposon strands are joined to 5’
overhanging
target strands exposed by a staggered
double-strand break at the target (Mizuuchi, 1984). However, Tn7 is cut away from the donor site by staggered
double-strand breaks, whereas cleavage of the 3’ and 5’
Mu strands occurs at very different stages, recombination
being initiated by single-strand breaks at 3’ transposon
ends. The exact positions of the 5’ Mu strand cleavages
remain to be established.
The movement of Tn7 is, in several respects, highly
reminiscent of the movement of the bacterial transposon
TnlO (Kleckner, 1989). TnlO also translocates via a nonreplicative mechanism (Bender and Kleckner, 1988) that
involves double-strand
breaks at the transposon termini
(Morisato and Kleckner, 1987) an excised transposon is
a key recombination intermediate (Haniford et al., 1991),
and chemically similar joints are formed between the
transposon ends and target DNA (Benjamin and Kleckner,
1989). The ends of Tnl 0, however, appear to be exposed
by flush cleavages (Benjamin and Kleckner, 1989), whereas
those that expose the Tn7 ends are staggered.
A striking feature of Tn7 transposition revealed by this
work is the highly concerted nature of this reaction. It will
be interesting to determine how recognition of a single site
in a bacterial chromosome is achieved and is communicated through multiple recombination
proteins to trigger
strand cleavages at the transposon ends and thereby provoke transposition.
Expsrfmental
Procedures
DNA Substrates
The donor p&mid
pEM-1 contains a 1.6 kb mini-Tn7 element flanked
by E. mli ch romosomal
sequences
unrelated to attTn7 (Arciszewska
et at., 1989). The mini-Tn7 element contains a 166 bp Tn7L segment
and a 199 bp Tn7R segment flanking a kanamycin
resistance
gene;
these end segments
contain the Tn7 cis-acting
transposition
Sequences. The terminal nucleottdes
of Tn7 are designated
Ll and Rl ,
the numbers increasing toward the middle of the element. The attTn7
target plaemid pKAG4-3
(McKown
et al., 1986) contains a 150 bp
segment that includes the sequences
necessary
for affTn7target
activity. Target DNA lacking affTn7 is Bluescript
pKS+ (Stratagene.
San
Diego,CA). pMIM-1 containsamini-Tn7elementwhoee
terminidirectly
adjoin Mlul restriction sites (see Figure 4). The mini-Tn7 element with
flanking Mlul and BamHl sttes was made by PCR amplification (below).
Fotlowing dttn
with BamHI, the amplified transposon
was ligated
into BamHldigested
Bluescript
pKB+. Prior to use, the transposon
was released by Mlul diiion
and gel purified using Gene Clean.
pLAl1 and pLA24 are afUn
plasmids that also contain Tn7R segments. To make pLAl1, the EcoRl DNA fragment containing
R199
from pLA1 (McKown et al., 1967) was inserted by blunt ligation into
Dralldigested
pRM2 (McKown et al., 1967); position R199 is adjacent
to the polylinker
Hindlll site. To make plA24, the EcoRl fragment
containing Tn7R42 from pCW5 (Arciszewska
et al., 1989) was inserted
by blunt ligation into Dralldigested
pRM2; R42 is adjacent to the polylinker Hindlll site.
tns Strains
Protein fractions were obtained from E. coli strains containing various
tns plasmids. The bacterial strains used were: MC4100, F- araD739A
A(afgf~c)U769rpsLl5OrelAl
flbB5301 deoC1 ptsF25rbsR(Casadaban, 1976); NLC51, MC4100 vaP recA56 (McKown et al., 1987); and
SY 903.1, recA1 sti:TnlO,
A(/acgfo)
ergEarn araD RifR NaN F’lecP’
lacZ::Tn5
(Sauer et al., 1988). The tnsA strain was NLCBt carrying
pKAO52 (Orle and Craig, 1990) the fns6 strain was SY 903.1 carrying
ptac-ms8
(Arciszewska
et al., 1989) the trrsC strain was NLCCI carrying pKAO53 (Orle and Craig, 1990) and the tnsDstrain
was MC41 00
carrying pCW23 (Waddell and Craig, 1966).
Cell Growth
and Pmparation
of tns Fractions
Cells were grown at 37OC in LB broth supplemented
with 100 pg/
ml carbenicillin.
Cells were grown to an OD,
of approximately
0.6,
harvested
by centrifugation.
washed in buffer A (25 mM HEPES [pH
7.51. 1 mM EDTA, 2 mM dithiothreitol
[DTtJ, and 100 mM KCI) (Fuller
et al., 1981) and frozen as cell paste. Crude cell lysates (stage I) were
prepared
by a freeze-thaw
lysis method (Fuller et al., 198t) using
buffer A. The protein concentration
of all lysates was about 20 mglml
except for that of NLC51 pKAO53, which was about 10 mglml. Stage
II tnsA and tnsCfractions
were prepared by treatment of the crude cell
lysates with 1% polyethylenimine
in buffer A at 509 mM KCI. collection
of the resulting supernatant
bycentrifugation,
protein precipitation
with
69% saturation
ammonium sulfate, and resuspension
and dialysis of
the resulting pellet with 25 mM HEPES (pH 7.5) 1 mM EDTA, 2 mM
DTT, and 500 mM KCI. The fns6 fraction was fraction II, about 90%
TnsB polypeptide
(L. Arciszewska,
R. McKown, and N. L. C., unpub
lished data). The tnsD fraction was stage Ill (about 1% TnsD, K. Kubo
and N. L.C.. unpublished
data). The host extract fraction was prepared
by boiling NLC51 crude cell lysate for 2 min and collecting the resulting
supernatant
after centrifugation
in a microfuge for 2 min at room temperature.
In Vitro Tmnaposltion
Reactions
Unless otherwise
indicated, all reaction mixtures (100 pl) contained
0.1 pg (0.25 nM) of pEM donor plasmid. 2.0 ug (10.0 nM) of pKAG43 affTn7 target plasmid, 2 mM ATP, 26 mM HEPES (pH 8.0). 1.3 mM
Tris-HCI
(pH 8.0) 2.5 mM KPG,, 0.1 mM EDTA, 2.2 mM DlT, 15 mM
MgAc, 100 pglml tRNA, 50 pg/ml SSA, and 5% polyvinyl alcohol MW
8000. Reactions in Figures 2B,B, and 7 also contained approximately
10 pg of stage II tnsA extract, 190 ng of TnsB (fraction II), 2.5 pg of
stage II msC extract, 4 pg of stage Ill fnsD fraction. 1 pg of stage II host
extract, 128 mM KCI. 15 mM NaCI, and 0.8% glycerol (v/v). Reactions
in Figures 2A, 3, and 48 contained 5 pg of stage II tnsA fraction, 75
ng of TnsB, 2.5 pg of stage II fnsC fraction, 4 pg of stage Ill msD
fraction, and 1 pg of host fraction; the reactiins
in Figures 2A and 48
also contained 50 mM KCI and 6 mM NaCI, and those in Figure 3 also
contained
136 mM KCI and 2.6 mM NaCI. The reactions in Figure 4A
contained 80 pg of stage I tnsA fraction, 75 ng of TnsB, 15 pg of stage
I msC fraction, 4 pg of stage Ill msD fraction, 36 mM KCI, and 3 mM
NaCI, in addllion to the NaCl indicated in the figure. The reactions in
Figure 9 contained a lower concentration
(3 nM) of target plasmid, 40
pg of stage I tn.sA fraction,
109 ng of TnsB, 10 pg of stage I msC
fraction, 4 pg of stage Ill msD fraction, 128 mM KCI, 15 mM NaCI, and
0.8% (v/v) glycerol.
The order of addition of the components
is unimportant,
except for
MgAc, which must be added after preincubation
of the other components for 7 min at 39OC. After precincubation,
MgAc is added, and
incubation is continued for 30 min at 30°C. MgC& can be substituted
for MgAc, but neither CaCI? nor spermidine is effective. Reactions are
stopped and the DNAs recovered
by a modification
of the Hoopes
protocol (Hoopes
and McClure,
1981; G. Glcor and G. Chaconas,
personal communication).
Four hundred microliters
of 10 mM TrisHCI (pH 7.5). 100 mM KCI. 5 mM EDTA, and 55% (w/v) urea are added
to the reactions,
followed by 66 sl of 100 mM spermine and then
Tn7 Transposition
815
In Vitro
incubation for 15 min on ice. The resulting pellet iscollected
by centrifugation, resuspended
in 200 pl of 10 mM MgAc and 300 mM NaAc,
tRNA added (150 pglml), and the mixtures incubated for IO min on ice.
DNA is recovered
by ethanol precipitation
and resuspended
in 25 ul
of 10 mM Tris-HCI
(pH 8.0) 1 mM EDTA, and 10 pg/ml RNAase.
Restriction
and Gel Electrophoresls
Reaction DNA(5 pl) is digested with 30-40 U
for 30 min at 37OC. (These rapid digestion
cause of nucleases in the extracted
DNAs.)
are electrophoresed
in 0.7% agarose gels
buffer.
of EcoRl in 10 ul reactions
conditions were used beAfter digestion, the DNAs
in 1 x Tris-borate-EDTA
Southern
Hybridlzatlon
DNA was electrotransferred
to Nytran
membranes
(0.45 micron,
Schleicher and Schuell), and the DNA is covalently bound by ultraviolet
irradiation in a Stratalinker
(Stratagene).
The membrane is probed with
a radioactive
DNA fragment specific
for the mini-Tn7 element-the
1200 bp Pstl fragmentgenerated by random priming labeling (Feinberg and Vogelstein,
1984) using [a-=P]dATP
and the Klenow fragment of DNA polymerase
1. Hybridizations
are carried out at 70°C
in a Hybrid-Eaze
chamber (Hoefer Scientific Instruhtents).
A l-3 hr
prehybridization
(1 .O M NaCI, 1 .O% SDS) was followed by overnight
hybridization
(10% dextran sulfate, 1 .O M NaCI, 1 .O% SDS, 100 ug/
ml denatured
salmon sperm DNA) with probe and followed by four
washes in high stringency
buffer (0.1 x SSPE, 1% SDS) over a 2 hr
period.
Isolation
of the Simple Insertion
Transposltlon
Product
The in vitro reaction product species were recovered
and pooled from
20 individual reactions. After EcoRl digestion, the DNA waselectrophoresed on a 0.7% agarose gel in 1 x Tris-acetate-EDTA
buffer. The gel
was stained with 0.5 pg/ml ethidium bromide, DNA was visualized with
long wave-length
ultraviolet light, gel slices containing
the desired
species were excised, and the DNA was purified with Gene Clean.
Ollgonucleotlde
Primers
The primers are complementary
to sequences
at the junctions of the
ends of Tn7 and atfTn7 (see Figure 5; sequences
described
in Craig,
1989). Tn7-L is complementary
to L31-60 of the top strand at the left
end of Tn7; Tn7-R is complementary
to RlOl-130
of the bottom strand
at the right end of Tn7. attTn7-L is complementary
to attTn7-25
to -5
on the bottom strand of aftTn7 and also has at its 5’ end the 5 nt
complementary
to the flanking polylinker
region in pKAO4-3,
which
flanks aftTn7 -25; attTn7-R is complementary
to atfTn7 +32 to +58 on
the top strand of affTn7 and also has at its Blend TTAA. Oligonucleotides were labeled at their Blends using T4 polynucleotide
kinase and
[y-“PJATP
(8000 Cilmmol; ICN Radionucleotides).
Primer Extensions
Extension reactions (10 J) contained approximately
0.2 pmol of endlabeled oligonucleotide,
approximately
10 ng of isolated simple insertion transposition
product, 1 U of Taq polymerase
(Cetus), dNTPs as
indicated below, and PCR buffer (10 mM Tris-HCI
[pH 8.31, 50 mM
KCI, 2.0 mM MgC12, 0.01% gelatin). In some reactions
(Figure 6)
dNTPs were added to 0.2 mM. Under these conditions,
Taq polymerase is highly processive,
but l-2 nt are added to the 3’ end of the
extension
product owing to a high level of nontemplated
nucleotide
addition (Clark, 1988). In other reactions (Figure 7) dNTPs were added
to 2 mM. Under this condition, no template-independent
addition is
detected on the 3’ end of the extension product (see control in Figure
7) but the extensions
are not highly processive.
The extension products and DNA sequencing
reactions, using the same primers as markers, were run on gels containing 6% polyacrylamide
(19:1 acrylamide:
bisacrylamide),
42% (w/v) urea, and 1 x Tris-borate-EDTA.
Productlon
of the Mlul Transposon
by PCR Amplification
We synthesized
an R-terminusoligonucleotide
primer complementary
to Rl-19
of the top strand at the right end of Tn7 and to 17 of 19
positions in Ll-19 of the bottom strand at the left end of Tn7 (positions
L14 and L16 differ from RI4 and R16), plus an additional Mlul site, a
BamHl site, and two extra nucleotides.
Using this oligonucleotide
and
a mini-Tn7 plasmid as template (Mullis and Faloona, 1987) PCR ampli-
fication
flanked
above).
pEM, 2
was used to generate a linear transposon
whose ends were
by Mlul and BamHI, which was used to produce pMIM-1 (see
Reactions contained
100 pmol of primer, 1 fmol of EcoRI-cut
mM dNTPs, and PCR buffer.
Production
of Control PCR Transposon
PCR amplification
and a mini-Tn7 plasmid template were used to produce a control transposon
whose 5’ends terminate at the last transposon nucleotide (Figure 7). The primer was complementary
to Rl-19 of
the top strand at the right end of Tn7 and to 17 of 19 positions in Ll19 of the bottom strand at the left end of Tn7 (positions L14 and L16
differ from RI4 and R16). The 5’ ends of this PCR transposon
are
provided by the primer, and the 3’ ends will be generated by polymerization; thus, the 5’ ends have a known structure.
The resulting PCR
transposon
was purifed through two cycles of agarose gel electrophoresis with Gene Clean extraction.
Acknowledgments
We thank other members of the laboratory
for materials and advice:
Karina Orle for tns plasmids and anti-Tns antibodies,
Lidia Arciszewska for TnsB, and Ken Kubo for TnsD fractions.
Glenn Gloor, George
Chaconas,
and Pat Brown also provided useful methods. We are especially grateful to Rudi Grosschedl,
Ann Hagemann,
Donald Morisato,
Anne Stellwagen, Evi Strauss, and Candy Waddell for their comments
on the manuscript.
R. B. was supported
by funds from the Medical
Scientist Training Program (NIH-MSTP-GM0
7618) Merck Sharp and
Dohme Research Laboratories,
the UCSF Department of Biochemistry
Tompkins
Memorial Fund, and the UCSF Program in Biological Sciences. P. G. was supported by funds from the Weingarten
Foundation
and the NIH-CNRS
program,
and by a NATO fellowship. The work
was supported
by a grant from the National Institute of Allergy and
Infectious Disease to N. L. C.
The costs of publication
of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked “advertisement”
in accordance
with USC 18 Section 1734
solely to indicate this fact.
Received
January
16, 1990; revised
March
21, 1991
References
Adzuma, K., and Mizuuchi, K. (1988).
tion reflects a differential distribution
266.
Target immunity of Mu transposiof Mu B protein. Cell 53, 257-
Arciszewska,
L. K., Drake, D., andcraig,
N. L. (1989). Transposon
Tn7
cis-acting
sequences
in transposition
and transposition
immunity. J.
Mol. Biol. 207, 35-52.
Barth, P. T., Datta, N., Hedges, R. W., and Grinter, N. J. (1976). Transposition of a deoxyribonucleic
acid sequence
encoding trimethoprim
and streptomycin
resistances
from R483 to other replicons. J. Bacteriol. 125, 800-810.
Bender, J., and Kleckner,
poses by a nonreplicative
N. (1986). Genetic evidence that TnlO transmechanism.
Cell 45, 801-815.
Benjamin, H. W., and Kleckner,
by TnlO. Cell 59, 373-383.
N. (1989).
Berg, D. E., and Howe, M. M., eds. (1989)
DC: American Society for Microbiology).
Intramolecular
transposltron
Mobile DNA (Washington
Bowerman, B., Brown, P. O., Bishop, J. M., and Varmus, H. E. (1989).
A nucleoprotein
complex mediates the integration of retroviral DNA.
Genes Dev. 3,469-478.
Brown,
Correct
P. O., Bowerman, B., Varmus, H. E., and Bishop, J. M. (1987).
integration
of retroviral DNA in vitro. Cell 49, 347-356.
Brown, P. O., Bowerman, B., Varmus, H. E., and Bishop, J. M. (1989).
Retroviral integration: structure
of the initial covalent product and its
precursor,
and a role for the viral IN protein. Pro?:. Natl. Acad. Sci. USA
86,2525-2529.
Casadaban,
M. J. (1976) Transposition
and fusion of the lac genes to
selected promoters
in Escherichia
coli using bacteriophage
lambda
and Mu. J. Mol. Biol. 704, 541-555.
Clark, J. M. (1988).
Novel non-templated
nucleotide
addition
reactions
Cdl
616
catalyzed
by procaryotic
Acids Res. 76.9677-9666.
and
Craig, N. L. (1989). Transposon
M. M. Howe, eds. (Washington,
ogy), pp. 21 l-225.
eucaryotic
DNA
Tn7. In Mobile
DC: American
polymerases.
Nucl.
DNA, D. E. Berg and
Society for Microbiol-
Mizuuchi, K. (1964). Mechanism of transposition
of bacteriophage
Mu:
polarity of the strand transfer reaction at the initiation of transposition.
Cell 39, 395404.
Morisato,
formation
D., and Kleckner,
N. (1967).
in vitro. Cell 51, 101-l 11.
TnlO
transposition
and circle
Craigie, Ft., and Mizuuchi, K. (1965). Mechanism
of transposition
of
bacteriophage
Mu: structure
of a transposition
intermediate.
Cell 47,
667-876.
Mullis, K. B., and Faloona, F. A. (1967). Specific synthesis
of DNA in
vitro via a polymerase-catalyzed
chain reaction. Meth. Enzymol. 755,
335-350.
Echols. H. (1990). Nucleoprotein
transcription,
and site-specific
14697-l 4700.
Orle. K., and Craig, N. L. (1990). Identification
of transposition
encoded by the bacterial transposon
Tn7. Gene 98, 1-7.
structures
initiating DNA replication,
recombination.
J. Biol. Chem. 265,
Eichinger,
D. J., and Boeke, J. D. (1966). The DNA intermediate
in
yeast Tyl element transposition
copurifies
with virus-like
particles:
cell-free Tyl transposition.
Cell 54. 955-966.
Eichinger,
is required
D. J., and Boeke, J. D. (1990). A specific terminal
for Ty transposition.
Genes Dev. 4, 324-330.
structure
Engels. W. Ft., Johnson-Schlitz,
D. M., Eggleston, W. B., and Sved,
J. (1990). High-frequency
P element loss in Drosophila
is homolog
dependent.
Cell 62, 515-525.
Feinberg, A. P., and Vogelstein,
B. (1964). A technique
for radiolabelling DNA restriction endonuclease
fragments
to high specific activity. Anal. Biochem. 137, 266-267.
Fujiwara,
structure
T., and Miruuchi,
K. (1966).
of an integration intermediate.
Retroviral
DNA integration:
Cell 84, 497-504.
Fuller, R. S.. Kaguni, J. M.. and K0rnberg.A.
(1961).
tion of the origin of the Escherichia
coli chromosome.
Sci. USA 78, 7370-7374.
EnzymaticreplicaProc. Natl. Acad.
Gay, N. J., Tubulewicz,
V. L., and Walker, J. E. (1966). Insertion of
transposon
Tn7 into the Escherichia
coli g/mS transcriptional
terminator. Biochem. J. 234, 11 l-l 17.
Gellert, M. K., Mizuuchi, K.. O’Dea, M. H.. and Nash, H. A. (1976). DNA
gyrase: an enzyme that introduces
superhelical
turns into DNA. Proc.
Natl. Acad. Bci. USA 73, 3872-3876.
Grindley, N. D., and Sherratt, D. J. (1976). Sequence
analysis
insertion sites: Models for transposition.
Cold Spring Harbor
Quant. Biol. 43, 1257-1261.
of IS1
Symp.
Gringaur,
E., Orle, K. A., Waddell, C. S.. and Craig, N. L. (1986).
Recognition
of Escherichia
coli attTn7 by transposon
Tn7: lack of specific sequence requirements
at the point of Tn7 insertion. J. Bacterial.
170,2832-2640.
Haniford, D. B., Benjamin, H. W., and Kleckner,
N. (1991). Kinetic
and structural
analysis of a cleaved donor intermediate
and a strand
transfer intermediate
in TnlO transposition.
Cell 84, 171-179.
Hauer, B., and Shapiro, J. A. (1984).
Gen. Genet. 194, 149-156.
Control
of Tn7 transposition.
Mol.
Hoopes, B. C., and McClure, W. R. (1961). Studies of the selectivity
for DNA precipitation
by spermine.
Nucl. Acids Res. 9, 5493-5504.
Kleckner, N. (1969). Transposon
M. M. Howe, eds. (Washington,
ogy), pp. 227-268.
Kleckner, N. (1990). Regulation
Cell Biol. 6, 297-327.
TnlO. In Mobile DNA, D. E. Berg and
DC: American
Society for Microbiolof transposition
in bacteria.
Annu. Rev.
Kubo, K., and Craig, N. L. (1990). Bacterial transposon
Tn7 utilizes two
different classes of target sites. J. Bacterial. 782, 2774-2776.
Lee, C. H., Bhagwat, A.. and Heffron,
F. (1963).
transposon
Tn3 sequence
required for transposition
Natl. Acad. Sci. USA 80, 6765-6759.
Identification
immunity.
Lichtenstein,
C., and Brenner, S. (1962). Unique insertion
in the E. coli chromosome.
Nature 297, 601-603.
of a
Proc.
site of Tn7
McKown,
R. L., Waddell,
C. S., Arciszewska,
L. A., and Craig,
N. L. (1967). Identification
of a transposon
Tnirdependent
DNAbinding activity that recognizes
the ends of Tn7. Proc. Natl. Acad. Bci.
USA 84, 7607-7611.
McKown, R. L., Orfe, K. A., Chen. T., and Craig, N. L. (1966). Sequence
requirements
of Escherichia
coli affTn7, a specific site of transposon
Tn7 insertion. J. Bacterial. 770. 352-356.
Mizuuchi,
chemical
K. (1963). In vitro transposition
approach to a novel replication
of bacteriophage
Mu: a bi*
reaction. Cell 35, 785-794.
Pato, M. L. (1969). Bacteriophage
M. M. Howe, eds. (Washington,
proteins
Mu. In Mobile DNA, D. E. Berg and
DC: American
Society for Microbiol-
ogy), PP. 23-52.
Qadri, M. I., Flares, C. C., Davis, A. J., and Lichtenstein,
C. P. (1989).
Genetic analysis of affTn7, the transposon
Tn7 attachment
site in Escherichia
coli, using a novel MlSbased
transduction
assay. J. Mol.
Biol. 207, 85-98.
Roberts, D., and Kleckner,
N. (1966). TnlO transposition
promotes
RecAdependent
induction of a X prophage. Proc. Natl. Acad. Sci. USA
85, 6037-6041.
Rogers, M., Ekaterinaki.
sis of Tn7 transposition.
N., Nimmo. E., and Sherratt, D. (1986).
Mol. Gen. Genet. 205, 550-556.
Analy
Sauer, R. T., Smith, D., and Johnson,
A. D. (1966). Flexibility of the
yeast alpha-2 repressor
enables it to occupy the ends of its operator,
leaving the center free. Genes Dev. 2, 807-816.
Shapiro, J.A. (1979). Molecular modelforthetranspositionand
tion of bacteriophage
Mu and other transposable
elements.
Acad. Sci. USA 78, 1933-1937.
replicaProc. Natl.
Thompson,
J. F., and Landy, A. (1969). Regulation of bacteriophage
lambda site-specific
recombination.
In Mobile DNA, D. E. Berg and
M. M. Howe, eds. (Washington,
DC: American Society for Microbiology), PP. l-22.
Varmus, H. E., and Brown, P. (1969). Retroviruses.
In Mobile DNA,
D. E. Berg and M. M. Howe, eds. (Washington,
DC: American Society
for Microbiology),
pp. 53-109.
Waddell, C. S., and Craig, N. L. (1988). Tn7 transposition:
two transposition pathways
directed by five Tnir-encoded
genes. Genes Dev. 2,
137-l 49.
Waddell, C. S., and Craig, N. L. (1989). Tn7 transposition:
recognition
of the affTn7 target sequence.
Proc. Natl. Acad. Sci. USA 86, 39583962.